A Study on Random Access Performance in Next Generation Mobile Network Systems

(1)

Institutionen för systemteknik

Department of Electrical Engineering

Examensarbete

A Study on Random Access Performance in Next

Generation Mobile Network Systems

Examensarbete utfört i Kommunikationssystem vid Tekniska högskolan vid Linköpings universitet

av Magnus Thalén LiTH-ISY-EX--15/4850--SE

Linköping 2015

Department of Electrical Engineering Linköpings tekniska högskola

Linköpings universitet Linköpings universitet

(2)

(3)

A Study on Random Access Performance in Next

Generation Mobile Network Systems

Examensarbete utfört i Kommunikationssystem

vid Tekniska högskolan vid Linköpings universitet

av

Magnus Thalén LiTH-ISY-EX--15/4850--SE

Handledare: Antonios Pitarokoilis

isy_{, Linköpings universitet}

Pål Frenger

Ericsson AB

Examinator: Danyo Danev

isy, Linköpings universitet

(4)

(5)

Avdelning, Institution Division, Department

Communication Systems

Department of Electrical Engineering SE-581 83 Linköping Datum Date 2015-07-02 Språk Language Svenska/Swedish Engelska/English Rapporttyp Report category Licentiatavhandling Examensarbete C-uppsats D-uppsats Övrig rapport

URL för elektronisk version

http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-120551 ISBN

— ISRN

LiTH-ISY-EX--15/4850--SE Serietitel och serienummer Title of series, numbering

ISSN —

Titel Title

En studie i random access prestanda i nästa generations mobila nätverkssystem A Study on Random Access Performance in Next Generation Mobile Network Systems

Författare Author

Magnus Thalén

Sammanfattning Abstract

The next generation of mobile telecommunication, 5G, will be specified in the near future. One of the proposed changes relative to the previous generation, 4G, is the inclusion of a new system control plane (SCP). The purpose of the SCP is to improve system scala-bility, forward compatiscala-bility, peak performance and to enable a higher degree of support for advanced antenna techniques. This is done by logically separating data transmitted explicitly from and to the user, the dynamic transmissions, from the broadcasted trans-missions that remain constant regardless of user activity, the static transtrans-missions, and by then redesigning the static part to make it more lean. This is expected to have several positive effects such as considerably more free resources, resulting in energy savings and potentially increased usage of MIMO. Knowing what effect the SCP has upon aspects such as random access is of importance when designing the solution that will go into the standard. Simulations show that there is potential in the inclusion of the new SCP. While the simulated 5G candidate systems that include the SCP have an increased delay when running the random access procedure, some aspects of the procedure have been improved. The main differences relative to the simulated 4G system are the performance of the first message in the procedure, which increased, and the performance of the second message in the procedure, which decreased. The deficiencies found in the handling of the second message, however, can be alleviated by using a more proper algorithm and channel design than what was used in this study.

Nyckelord

(6)

(7)

Sammanfattning

Nästa generationes mobila kommunikation, 5G, kommer specificeras inom en snar framtid. En av de föreslagna förändringarna relativt den tidigare genera-tionen, 4G, är tillägget av ett nytt systemkontrollplan. Syftet med detta system-kontrollplan är att förbättra systemets skalbarhet, framåtkompabilitet, maxpre-standa och att möjliggöra en högre grad av använande av avancerade anntenn-tekniker. Detta görs genom att logiskt separera den data som explicit sänds till och från användare, dynamisk transmission, från den data som sänds oavsett om det finns någon aktiv användare eller ej, statisk transmission, och genom att se-dan förändra den statiska delen så att den blir smalare. Detta förväntas ha flera positiva effekter så som avsevärt mer oanvända resurser vilket resulterar i ener-gibesparingar och potentiellt en högra grad av MIMO användade. Att veta vilka effekter systemkontrollplanet har på aspekter som random access-proceduren är av stor vikt när man ska designa den lösning som ska bli del av standarden. Simulationer visar att det finns potential i tillägget av systemkontrollplanet. De simulerade 5G kandidatsystemen som innehåller systemkontrollplanet har en ökad fördröjning vid använande av random access-proceduren, men vissa aspek-ter av proceduren har förbättrats. De huvudsakliga skillnaderna relativt det simu-lerade 4G systemet är prestandan av det första meddelandet i proceduren, som förbättrades, och det andra meddelandet i proceduren, som försämrades. Bris-terna i hanteringen av det andra meddelandet kan dock åtgärdas genom en mer lämplig algoritm och kanaldesign än vad som användes i denna studie.

(8)

(9)

Abstract

The next generation of mobile telecommunication, 5G, will be specified in the near future. One of the proposed changes relative to the previous generation, 4G, is the inclusion of a new system control plane (SCP). The purpose of the SCP is to improve system scalability, forward compatibility, peak performance and to en-able a higher degree of support for advanced antenna techniques. This is done by logically separating data transmitted explicitly from and to the user, the dynamic transmissions, from the broadcasted transmissions that remain constant regard-less of user activity, the static transmissions, and by then redesigning the static part to make it more lean. This is expected to have several positive effects such as considerably more free resources, resulting in energy savings and potentially increased usage of MIMO. Knowing what effect the SCP has upon aspects such as random access is of importance when designing the solution that will go into the standard.

Simulations show that there is potential in the inclusion of the new SCP. While the simulated 5G candidate systems that include the SCP have an increased de-lay when running the random access procedure, some aspects of the procedure have been improved. The main differences relative to the simulated 4G system are the performance of the first message in the procedure, which increased, and the performance of the second message in the procedure, which decreased. The deficiencies found in the handling of the second message, however, can be allevi-ated by using a more proper algorithm and channel design than what was used in this study.

(10)

(11)

Acknowledgments

First and foremost I would like to thank my supervisor at Ericsson AB, Pål Frenger, and my supervisor at Linköping University, Antonios Pitarokoilis, for their con-stant support throughout these past six months. It is thanks to you that I have been able to move forward despite facing many difficulties along the way. I also wish to thank everyone else who has been involved in my thesis for cre-ating a great environment that has made this a very positive experience for me. Your presence has been invaluable and has helped me greatly in realising this thesis.

Linköping, June 2015 Magnus Thalén

(12)

(13)

Notation

Abbreviations

Abbreviation Definition

3GPP 3rd Generation Partnership Project bps bits per second

CDF Cumulative Distribution Function dB deciBel

dBW deciBel Watt

EPS Evolved Packet System ISD Inter Site Distance LTE Long Term Evolution MAC Medium-Access Control

MBSFN Multi Broadcast Single Frequency Network MIMO Multiple Input Multiple Output

PDCCH Physical Downlink Control Channel PDCP Packet Data Convergence Protocol PDSCH Physical Downlink Shared Channel

PHY Physical Layer

PRACH Physical Random-Access Channel PUCCH Physical Uplink Control Channel PUSCH Physical Uplink Shared Channel

RAN Radio-Access Network RAR Random Access Response RLC Radio-Link Control SCP System Control Plane

(16)

(17)

1

Introduction

For the last few decades a new generation of mobile communication systems has been developed with a periodicity of about ten years. Next in line is the fifth generation, 5G, which is currently in the process of being specified. Due to the ever increasing demands combined with new research being implemented the ex-pectations are greater than ever. 5G is expected to deliver rates up to 10Gbps for up to a hundred times as many devices as today, with a latency of less than 1ms. Furthermore a 90 percent reduction in energy usage as well as virtually complete availability and worldwide coverage is expected [14].

In this thesis we aim to investigate a subset of changes that may potentially be implemented in the new standard. These modifications might fundamentally al-ter the design of contemporary communication systems. More specifically we are interested in how these changes affect the procedure used when mobile users perform the initial procedure to connect to the cellular network, i.e. the random access procedure.

1.1 Motivation

Due to their close relation to the telecommunication field, Ericsson as well as other major players have a large interest in participating in the ongoing standard-ization effort. Considering that 5G is intended to be out on the market beyond year 2020 [13] this is the time to get involved. It is believed that 5G will need to undergo fundamental changes, relative to previous generations, in order to reach the high expectations on data rates and latencies [8].

The predecessor of 5G, 4G, was a considerable upgrade relative to previous sys-tems. 4G contains functions such as spectrum flexibility, multi-antenna

(18)

2 1 Introduction

sion and intra-cell interference control [7]. All of this has made 4G a relevant technology, and considering the longevity of even earlier generations it will con-tinue to be so for many years to come.

All wireless transmissions make use of electromagnetic waves, meaning that the physical nature of the signals are electromagnetic waves. The nature of the elec-tromagnetic waves is such that an antenna receiving two different signals at the same time, where the signals are of the same frequency, will be unable to tell the signals apart. The receiver equipped with that antenna will only be able to per-ceive the sum of the signals.

In order to avoid excessive interference between signals 4G schedules all signals in a certain service area on different time and frequency resources. This improves the probability of successful reception of transmitted messages, but it also im-poses some constraints as it further limits the total number of transmissions that can be done in a certain time window.

The increasing amount of time-frequency resources spent on overhead in 4G has become an increasingly large issue. As result of this there will of course be less resources left to use for transmission of other data, but another concern is the impact it has on future development.

The main task of wireless systems such as 4G is to transfer information from transmitter to receiver. In order to accomplish this it is necessary to also transmit certain control information. The control information, which for instance consists of transmission schedule and system settings, can be considered to be overhead as it is only transmitted as a means to make the system work.

One important kind of control signal is the reference signal that each base sta-tion transmits in regular intervals. The reference signals are necessary because they allow the users to gain a measurement of the channel between user and base station. As all users need this information the base stations broadcast the refer-ence signals, meaning that it is sent in a way that allows all users in the vicinity to receive it.

While there is a need of broadcasting reference signals it is not necessarily needed to constantly spend as many resources on this as is done in 4G. This can be con-sidered to be an improvement area as this causes the system to be hard to extend, and attempting to put in even more functionality would likely become an increas-ingly difficult and complicated process. Instead it would probably be beneficial to start over from a clean slate and reduce the amount of overhead clogging the system. Doing so would also make the system more energy efficient and reduce interference. As 5G is expected to evolve and change considerably compared to 4G, now is a good opportunity to implement this change.

(19)

1.1 Motivation 3

provide coverage to the users through strategically placed base stations. Base sta-tions are essentially towers equipped with antennas, enabling them to transmit and receive electromagnetic waves. The base stations are generally evenly dis-tributed as a single base station can provide coverage to an area around it. In places with more traffic the base stations are more densely deployed as a single base station is limited in the amount of traffic it can handle.

In LTE the term eNodeB is used to describe a base station. Each eNodeB con-sists of one or more cells. Each cell is an independent logical entity that serves a certain area around the eNodeB. All communication to and from all users in this area is handled by the designated cell.

When a user wants to connect to the system the user will try to perform random access to the cell that governs the area that the user is located in. As the name implies random access is the procedure when a user, that the system has no prior knowledge of, suddenly attempts to establish a contact with the system. In order for the user to be able to do this it needs to have received a reference signal and system information from the cell. In LTE the reference signals are unique to each cell, meaning that a reference signal obtained from one cell is only useful when trying to communicate with that specific cell. This results in the user being able to differentiate between cells, the user will realize if it moves into another cell as that cell would then transmit a different reference signal. The opposite of this would be a system where all reference signals were equal. The user would then have no way to distinguish between different cells. The desirable and flexible ap-proach would be to make both of these alternatives possible. This would allow the system to choose whichever reference signal setup that it finds most suitable at the time.

Enabling the user to make use of several cells in the random access procedure, as opposed to only one cell in LTE, has been identified as a promising change. If this was implemented, and all cells in an area broadcasted the same reference sig-nals, it would reduce interference between cells and thus allow for a more dense placement of eNodeBs without increasing system overhead.

An addition to 5G that has been suggested is the inclusion of a new system control plane (SCP) [4][5][6]. At its core the SCP is a subset of all transmitted data in the system. The SCP represents the data that is necessary for basic system function-ality, such as initial access to the network. This includes signals sent even when all users are idle, or when there are no users in the system, as well as signals the random access and paging procedures. This includes signals such as reference signals and some broadcasted system information, but it does not include data that is explicitly requested by the user. The purpose of this system control plane is specifically to reduce the amount of signals that are unnecessarily broadcasted while also logically separating the signals that are part of the SCP from the other signals, making it possible to configure the system so that users will be unaware of specific cells during the initial access to the system. As the problems are

(20)

in-4 1 Introduction

herent to the nature of the previous generations it would not be simple to solve through a new release of the LTE-standard. Instead it is convenient to include it as a more integrated change for the upcoming 5G.

The inclusion of the system control plane would change many parts of the sys-tem. The focus of this thesis, however, is specifically how it affects the random access procedures. Before the upcoming standardization effort it is important to fully understand the consequences of different suggestions. Having data that support your suggestions is key. Likewise it is important to be aware of any po-tential weaknesses of your suggestions, which is why it is crucial to gain insight into what effects the SCP has on random access, among other things.

1.2 Purpose

The aim of this thesis is to find to what extent the SCP is positive or negative to the random access procedures of the upcoming 5G system. To be able to make such a judgement we will need to look at different metrics that determine the effectiveness of the random access.

When performing the random access procedure a user will send and receive a certain set of messages in accordance with a protocol. The resulting signal to noise ratios (SNR) at the receiver of these messages as well as energy efficiency will need to be considered. The total time it takes for a user to connect to the sys-tem and the amount of concurrent users the syssys-tem can serve is also of interest. Depending on the achieved result it may be necessary to include additional changes to the random access procedure in 5G. Otherwise the inclusion of the SCP may cause the performance of the this procedure to deteriorate below acceptable lev-els. It is therefore necessary to gather data of performance of the random access functions during different circumstances and compare it to LTE. LTE is consid-ered as a baseline because it is the closest predecessor of 5G and is thus the clos-est to 5G in both performance and structure.

If the final results are positive it becomes more likely that the SCP, as described in this report will, be a part of 5G. While the SCP is considered to be a good change overall it is not certain that it will not negatively impact some parts of the system, making it necessary to investigate its effects on procedures such as random access.

1.3 Problem Formulation

Researchers have identified the inclusion of a new system control plane to be a promising change for the upcoming 5G [4][5][6]. While it is known that this would have benefits to some aspects of the system compared to the current LTE-standard, it is also possible that it will incur some drawbacks. The random access

(21)

1.4 Limitations 5 procedure is one of the areas that will be directly affected by this change. There-fore it is necessary to investigate in detail how metrics such as the delay and power consumption of the procedure are affected and how the procedure can be improved as a result of the SCP. The system’s capacity in terms of random ac-cess user throughput should also be considered. This thesis aims to answer the following:

• How is the power control for the random access procedure affected by the SCP?

• How does the SCP affect the random access procedure in terms of delay compared to the 4G baseline?

• How can the random access procedure that uses the SCP be improved and how do these changes affect the delay and user throughput of the proce-dure?

The first question refers to the power that users estimate that they will need to use in order to successfully transmit the preamble to its intended receiver. The result will differ from 4G as the SCP enables the user to combine reference sig-nals from several different cells.

The SCP is certain to affect the performance of the random access procedure, though it is not known whether the effect will be positive or negative. The pur-pose of the second question is to gain information on this matter.

In the event that the SCP negatively impacts the random access procedure it may be necessary to find ways to decrease this effect. The third question refers to any potential changes that will improve the procedure. In this case it is mainly the performance in terms of total delay and user throughput that is considered.

1.4 Limitations

When analysing the capabilities of the system there are different parameters that can be changed. Each possible configuration is likely to yield quite different re-sults. It is not possible to investigate all of the possible circumstances relevant to the random access procedures, so this thesis is limited to a smaller set of rep-resentative cases. It is of interest to see how performance metrics, such as delay, are affected by changes in some key variables. Except for these few variables, essentially everything will be held constant. The variable parameters are:

• Inter site distance, the distance between neighbouring eNodeBs.

• Number of preambles, the number of preambles that users may choose from when transmitting the first message of the random access procedure. Each user chooses one.

(22)

6 1 Introduction

• Intensity, the number of users/s that are added to the system and that the system is tasked to service.

5G will of course consist of many changes when compared to LTE, but the only change considered in this report is the inclusion of the SCP. Therefore the 5G test system discussed in this report is actually close to LTE in functionality. Most other changes that are likely to be included in 5G will probably not significantly impact the random access procedure and are therefore not included in the 5G test system used in this report.

1.5 Thesis Outline

This report is divided into an introductory part that presents some theory and background, and a result part that covers the execution and achieved results.

• Chapter 1 gives an overview of the problem and the suggested solution, and presents the problem formulation.

• Chapter 2 introduces the concept of wireless transmission and some related technologies and phenomena.

• Chapter 3 presents the typical structure of cellular networks and provides information about the cellular networks system named 4G. Some informa-tion about its structure and protocols are described. Addiinforma-tionally the ran-dom access procedures used in 4G are described in more detail.

• Chapter 4 gives and introduction to the SCP. The issues that have prompted the inclusion of the SCP into future cellular network systems are presented. Furthermore the actual changes that constitute the SCP and the expected benefits are described.

• Chapter 5 details the scenario as well as parameters used in the simulator during the different stages of the investigation.

• Chapter 6 presents the results.

• Chapter 7 discusses noteworthy results as well as possible weaknesses of the used method. The impact the thesis might have on society is briefly mentioned.

• Chapter 8 states what can be concluded from the results and attempts to answer the questions posed in the problem formulation. Suggestions of what additional research could be made in order to answer these questions are also made.

(23)

2

Wireless Transmission Overview

Wireless mobile systems make use of many different kinds of techniques to ac-quire sufficient quality of signal transmission. These techniques are essential for the level of modern cellular services. Further exploitation of these techniques is therefore something that is desirable in upcoming generations as well as in new releases of current mobile telephony systems. This is thus something that should also be kept in mind when analysing the topic of this report, the performance of the random access procedure in a future system. For this reason this chapter introduces the most relevant techniques and phenomena in order to give a frame-work of understanding that is helpful when weighing the options of 4G and 5G against each other.

This chapter briefly introduces some wireless transmission basics, fading and di-versity.

2.1 Basics

Typically when talking about wireless transmissions we are interested in the com-munication between two units, a transmitter and a receiver. The transmitter is the unit that transmits the signal while the receiver is the unit that receives the signal. The transmitter and the receiver do not necessarily have any special prop-erties that differentiates them, rather they are only defined by the direction the signal is going in. The signal always originates at the transmitter. The task of the receiver is to detect the signal and interpret it appropriately.

The physical entities that constitute the signals are virtually always electromag-netic waves. Electromagelectromag-netic waves travel at the speed of light and have many different names depending on the frequency of the wave, and by extension the

(24)

8 2 Wireless Transmission Overview

application in which it is used. For example, visible light, radio waves and mi-crowaves are all instances of electromagnetic waves. It is often convenient to model the electromagnetic waves as sinusoidal waves. The general equation

x(t) = a sin t, (2.1)

where x is the electromagnetic wave, a is the amplitude and t is time, can be used as a simplified model of an electromagnetic wave.

An important property of electromagnetic waves is that they decrease in power the further they travel. This shows through the amplitude which decreases and eventually becomes so small that the wave is no longer measurable at the receiver. The path loss is a metric that describes the attenuation of the signal strength. In free space conditions, where there are no objects obstructing the signal, the path loss can be calculated as

P = 4πd λ

!2

, (2.2)

where P is the path loss, d is the distance between the transmitter and the receiver and λ is the wavelength of the carrier of the transmitted signal. Because of this fundamental constraint the distance between the transmitter and the receiver is a limiting factor in wireless communication. Communication is only possible if the transmitter and the receiver are not separated by too large of a distance as in that case it will be impossible for the receiver to detect the transmitted signal due to its low energy.

If the only thing obstructing wireless communication was the path loss it might still be possible to communicate regardless of how small the amplitude of the transmitted signals are, as long as it is larger than zero. In practice however, there exist several other obstacles that prevent this. The perhaps most prevalent problem hindering this is the ubiquitous noise. Noise is a general term that is usu-ally used to describe all kinds of distortion of signals, except for the interference from other signals, and can have a number of causes. The most common kind of noise is that which is called white noise. White noise, which is a synonym to thermal noise, exists in all kinds of circuitry and also afflicts all electromagnetic waves that travel through the air. The white noise has a constant power spectral density, meaning that it is equally strong at all frequencies, and is generally mod-elled to have a Gaussian distribution.

If a signal at the transmitter can be described by (2.1) then the same signal at the receiver can in some simple cases be described as

y(t) = b sin (t + φ) + n(t), (2.3)

where y(t) is the received signal, b is the amplitude of the received signal, t is the time, φ is the phase shift relative to the transmitted signal and n(t) is the white noise. The receiving amplitude b can be calculated as

b = √a

(25)

2.1 Basics 9

where P is the path loss according to (2.2) and a is the amplitude of the sent signal at the transmitter. The noise n(t) on the other hand is not affected by the distance. This means that when a signal modelled as (2.3) has travelled far enough the noise portion of the signal will be significantly larger than the intended message, making it impossible for the receiver to make any sense of the signal. Figure 2.1 shows a sine wave before passing through a channel as well as what happens when the signal is attenuated and has noise added to it.

Time Amplitude Time Amplitude Time Amplitude

Figure 2.1:The effects on a sine wave when passing through a channel that attenuates the signal and applies additive white Gaussian noise.

When discussing the difference between the signal at the transmitter and the same signal at the receiver it is common to make use of something called the channel. The channel is essentially a filter between the transmitter and the re-ceiver through which the signal propagates. It describes exactly how the signal is affected by travelling through the air. The channel is dependent on the positions of the transmitter and the receiver, and the environment. Assuming that the envi-ronment changes, which is always the case in real scenarios, it is also dependent on the time. The channel is not, however, dependent on the signal that travels through it, all signals are treated the same way. A general relation between the transmitted and the received signal is

y(t) = h(t) ∗ x(t) + n(t), (2.5)

where y(t) is the received signal, h(t) is a mathematical representation of the channel, x(t) is the transmitted signal, n(t) is the noise, t is the time and ∗ de-notes convolution. The majority of wireless communication likely occurs close to the surface of the earth. The average call between two mobile phones, for example, will usually have the two phones at ground level. An effect of this is that the transmitted signals are obstructed by objects in the environment. For example buildings and mountains may very well block the path of the signal. Electromagnetic waves can not always pass through house walls, and certainly not mountains. What they can do is reflect and refract off objects. For this reason the transmitted signal will typically take several paths and bounce off different

(26)

objects before arriving at the receiver. This is called multipath propagation. Figure 2.2 shows the concept of multipath propagation. Two of the lines dis-played in the graph correspond to the two signals that arrive at the receiver after having travelled along two different paths. The third line, the signal which the receiver preceivs, is simply the sum of the two other signals.

Time

Amplitude

Signal at receiver Path 1

Path 2

Figure 2.2:An example of multipath propagation without noise.

2.2 Pre- and post-processing

Except for the actual transmission of the signal there are also some steps that, to some degree, need to be taken to prepare the signal for the distortive effects of the channel. This includes source coding, channel coding and modulating the message. The result is a compressed, error protected message that is ready to be mapped onto a waveform. After receiving the signal the receiver needs to reverse the operations in order to retrieve the original message. This typical flow is showed in Figure 2.3.

2.2.1 Source coding

The purpose of source coding is to compress and reduce the size of the target information. This is not something unique to wireless transmission but is rather done in very many different contexts. Many different source coding schemes exist and they yield a different degree of compression depending on the information that is to be compressed.

(27)

2.2 Pre- and post-processing 11 Channel coding Modulation Channel decoding Demodulation Source decoding Source coding Air interface Transmitter Receiver

Figure 2.3:A common order of operations for information that is to be trans-mitted from a transmitter to a receiver.

In wireless transmission the source coding is beneficial as it allows the transmit-ter to actually transmit shortransmit-ter bit sequences while losing little or no information. This comes at a computational cost both at the transmitter and at the receiver, but this is acceptable as it typically is the actual wireless transmission that is the bot-tleneck and not the pre- or post-processing.

2.2.2 Channel coding

Channel coding is included in order to counter errors that may appear in the re-ceiver due to distortion of the transmitted signal. When the rere-ceiver receives the signal it will first demodulate it, resulting in a series of bits. If not for the error correction it would be impossible to tell whether the information had been cor-rupted or not. With error correction, however, it may be possible to detect and possibly even correct a number of errors. In a practical situation this allows the receiver to either correct the errors or at least discard it and request a retransmis-sion of the message if needed.

The error correction operates on a set of bits and works by adding redundancy. The redundant bits are used to create a structure that allows the receiver to spot errors if they exist.

2.2.3 Modulation

The modulation is used in order to convert the message consisting of a series of bits into a waveform that can be transmitted over the air. The modulation essentially maps the bits to different looking sinusoids. The resulting signal that will then be transmitted is all the sinusoids appended in the same order as the bits that made up the message.

(28)

2.3 Fading

If a transmission between transmitter and receiver takes place in an area with different obstacles spread out, as opposed to in a free space, the signal will prop-agate towards the receiver along several paths as described in Section 2.1. In the end this causes the receiver to receive the summation of several distorted copies of the signal. Due to the distortion it is not certain that the received signal will look anything like the original transmitted signal. Signal properties such as the amplitude, phase and angle of arrival are subject to variations due to the sur-rounding environment [12]. One effect of this is that the power of the resulting signal at the receiver may at times be enhanced and at times reduced.

The fading phenomenon is traditionally divided into two parts depending on their rate of change: large-scale fading and small-scale fading. Large-scale fad-ing is the fadfad-ing component caused by large objects obstructfad-ing the signal and large movements. Small-scale fading on the other hand, which is also called Rayleigh fading [12], corresponds to the significant effects that can be caused by movements of the receiver as little as half a wavelength of the transmitted sig-nal. The name Rayleigh fading comes from the fact that the signal amplitude is distributed according to a Rayleigh distribution [12].

2.4 Diversity

Diversity is the method of providing more instances of the same information to a receiver in order to reduce the negative effect of fading. Fading, which is caused by multipath propagation, can randomly both improve and worsen the quality of a signal at the receiver. What happens when some kind of diversity is included is that the receiver, through for instance multipath propagation, receives several copies of the same information. Even if the receptions of each single instance is no better than before it becomes much more likely that at least one of the instances is of sufficient quality for the receiver to interpret correctly. Several dif-ferent ways to provide diversity exists.

2.4.1 Spatial Diversity

Spatial diversity is simply the method of attaining diversity by placing transmit-ters and receivers at different positions. Between each existing transmitter and receiver there will be at least one path. Therefore, as long as there is more than just one transmitter and one receiver there will be a gain in diversity. It is as-sumed that all transmitters send the same signal and that the receivers have a way of synchronising their information. However, this diversity gain can only be said to happen under a certain constraint. Only when the additional transmit-ters are spaced sufficiently far apart from the original transmitter, or when the additional receiver is spaced far enough apart from the original receiver. If, for

(29)

2.4 Diversity 13

instance, two receivers are located at almost the exact same position it is likely that the signals transmitted from the transmitters will propagate along the same paths. Such a case will not give additional diversity. The distance needed in order to gain diversity varies, but is at least 0.5λ, where λ is the wavelength of the transmitted signal [9]. A metric of how similar two units are in this sense is correlation. When the distance between two units increase the correlation will generally decrease.

One practical way of creating spatial diversity in cellular systems is by using several different system nodes to communicate with a user when sending a mes-sage. This method may be especially relevant to this thesis as the system control plane may allow this behaviour for certain procedures.

2.4.2 Delay Diversity

Delay diversity makes use of time differences in order to gain diversity. If a chan-nel is time dispersive the different propagation paths will cause different delays. In such a case the receiver will receive several instances of the transmitted signal, although they are probably corrupted and interfere with each other. This is an example of delay diversity. Note that this is possible even in the case where there is only a single transmitter and a single receiver. It relies on the quality of the channel. Even if the channel is not time dispersive it is possible to create artifi-cial time dispersion by transmitting the same signal from several antennas with small time delays in between [9]. If the antennas are placed apart this can also be seen as spatial diversity. The time dispersion and delay diversity can, however, also be achieved this way even when antennas are placed right next to each other, with less space between them than what is required for spatial diversity.

(30)

(31)

3

LTE Overview

Long-Term Evolution (LTE) is a mobile telephony system that was first publically available during 2009 [2]. It is a system consisting of logical units called eNodeBs and mobile users such as mobile phones. A traditional implementation of the eN-odeB is a base station. An eNeN-odeB consists of one or more cells, where a cell is also a logical construct. A single cell represents the interface that is used for transmission towards mobile phones in a specific area. This area that the logical cell governs is what is usually indicated when the word cell is mentioned. The eNodeBs are placed in such a way that they cover the area which they are sup-posed to provide coverage to. A typical, albeit idealistic, deployment of eNodeBs that is common for mobile telephony systems, is to arrange them in a hexagonal grid such that each eNodeB is located at the point where three hexagons intersect, see Figure 3.1.

Mobile users will typically communicate wirelessly with one of the cells of the eNodeB that is adjacent to whatever hexagon the user happens to be in at the moment. To accomplish this different techniques are used for transmission and reception of the signals that contain the requested data.

This chapter will give a short overview of LTE as a whole and will additionally go into more detail in the parts of LTE that are especially relevant to this thesis: the random access procedures as well as the parts of LTE necessary for random access.

(32)

16 3 LTE Overview

Figure 3.1: A hexagon deployment using three cells per eNodeB. The hexagons with the same pattern are the areas that are governed by the cells that belong to the closest eNodeB.

3.1 Cellular Networks Infrastructure

The most basic component of today’s cellular network systems can be considered to be the user. The users are typically very spread out and most of them are likely to move at some point. This causes the need for a widespread coverage. The pur-pose of the system is to provide services, such as the ability to call other users or browse the Internet, and these services should preferably be available at all times. The high requirements on availability require the deployment of certain infras-tructure. The basic unit that provides the coverage is the eNodeB. The eNodeB is a stationary unit that on one hand wirelessly communicates with the surround-ing users, and on the other hand communicates, typically through wires, with the underlying system. The main wireless part of cellular network systems is the part between the user and the eNodeB. The communication between eNodeBs are typ-ically done using a wired approach, or sometimes through dedicated radio links that do not interfere with other communication.

It would be ideal if each station provided as much coverage as possible, mean-ing that each eNodeB should be able to pick up wireless transmissions sent from an area as large as possible. Thus, if carefully positioned it would be possible to deploy less eNodeBs without affecting system performance, allowing the opera-tors to save money. A simple way of improving the eNodeBs ability to pick up

(33)

3.2 LTE Structure 17

signals is to give the eNodeBs antennas as clear a view as possible of the surround-ings. As mentioned in Chapter 2, physical objects have the ability to obstruct the signals, causing a degradation in signal quality. The placement of eNodeBs on high buildings is advantageous as there will on average be much fewer objects that obstruct the view between the eNodeB and the different nearby users. In addition to this it is of course also important to place the eNodeBs in such a way that the coverage they provide do not overlap unnecessarily. What limits the coverage is, except for obstructing objects, the distance between the eNodeBs and users. Assuming that measures have been taken to remove or lessen the issue of obstructing objects, the provided coverage can in some cases roughly be mod-elled as a circle with the eNodeB in the middle. A drawback of using circles to model the coverage area is, however, that it is very unwieldy to cover a large area completely with only circles without at the same time causing a lot of overlap be-tween them. A more simple way of fulfilling both demands of complete coverage and as little overlap as possible is to use hexagons as building blocks instead of circles.

Commonly the eNodeBs employ several sets of antennas placed around the eN-odeB, facing different directions. In such a case the antennas will not transmit in all directions, but they rather transmit their respective signals in an area spanned by an angle of 60 to 120 degrees, depending on the scenario. A way to structure the different antennas is to group all antennas facing a certain direction into one group. A common division of antennas is to divide them into three groups where each group of antennas transmits their respective signals to a third of the sur-rounding area, corresponding to a width of 120 degrees. In such a case each group of antennas could be mapped to the cell serving the corresponding area. A typical placement of eNodeBs consisting of three cells are shown in Figure 3.1. The result of this division is that the different groups all transmit their signals to essentially non-overlapping areas.

3.2 LTE Structure

LTE is at the topmost level divided into two parts: the Radio-Access Network (RAN) and the Evolved Packet System (EPS). The RAN takes care of all activities that are needed for a typical radio transmission between eNodeB and user such as scheduling, encoding of the signal, adding error correction and so on. The EPS, on the other hand, handles all other functionality that is needed to connect all entities in the network and make it run properly such as access to Internet, connection between eNodeBs as well as other necessary back end functions. It is primarily the RAN that is of interest in this thesis due to the random access being a part of it.

The RAN can be divided into several different protocol layers, as shown by Fig-ure 3.2, where a protocol layer contains functions necessary for transmission and

(34)

18 3 LTE Overview

reception of messages. Adjacent layers also provide interfaces for each other, al-lowing them to call on each other in order to accomplish their tasks. These layers range from the topmost layer that performs the initial processing of any intended message, to the physical layer which interfaces the physical antenna components. The layers fulfil different tasks and provide proper building blocks for the system so that the interfaces in between the layers are effectively designed. The different layers are the Packet Data Convergence Protocol (PDCP), the Radio-Link Control (RLC), the Medium-Access Control (MAC) and the Physical Layer (PHY).

MAC PHY MAC PHY RLC RLC eNodeB User PDCP PDCP

Figure 3.2: The different protocol layers of the RAN. Users and eNodeBs communicate with each other through the physical layer.

For each transmission there is a transmitter and one or more receivers. The differ-ent protocol layers exist in both the transmitter and the receiver but the related functions differ somewhat. Some functions, such as scheduling, are only per-formed for the network while the user simply adapts. The most common case, however, is that each function in a certain protocol layer at the transmitter has a mirroring function in the same protocol layer in the receiver. This is the case for functions such as encoding and modulation of signals which are mirrored by corresponding decoding and demodulation. It should also be mentioned that the behaviour of the layers in some cases depend on whether the message that is to be transmitted is on the downlink or on the uplink. A message is sent through the downlink if the message originated from an eNodeB and is sent to a user, and vice versa for the uplink. The difference is due to the fact that the demands on the preprocessing are different in the two cases. When sending from a user, on the uplink, the user has likely already received instructions from the eNodeB on when and how to send it.

(35)

3.2 LTE Structure 19

Messages that are to be sent using the RAN are in the form of packets called IP-packets. An IP-packet consists of a header and a payload part. The header is simply some necessary meta data describing the message, and the payload is the actual data being sent. Before and after passing through the air interface this message is altered according to the instructions of the different protocol layers. The first purpose of the PDCP is to compress the header of the packet in order to reduce the size of the message. The second purpose is to solve the privacy issue that is apparent when transmitting data over the air. After all, the transmitted signal will typically propagate with similar attenuation in all directions covered by the transmitting cell. To solve this problem the message is encrypted so that only the designated receiver is able to decipher it.

Upon receiving the compressed data blocks from the PDCP the RLC has the task of combining these blocks into larger or smaller blocks when necessary. The new data blocks are created by combining the incoming data blocks and splitting them where needed in order for the resulting data block to be of a specific length. The desired length can vary and depends on the data rate used, for high rates it is more efficient with large blocks whereas lower rates require smaller blocks. On the receiver side the RLC should also ensure that there are no errors in the mes-sages that will be passed to the layer above it, the PDCP. If it detects any errors it is able to request a retransmission.

The MAC defines two sets of channels, the logical channels and the transport channels. Both of these kinds of channels are abstract in the sense that no sig-nals are sent through the air; instead the channels function as interfaces to the surrounding protocol layers. The logical channels act as an interface between the MAC and the RLC and are defined by the type of information that is to pass through the channel. The transport channels instead act as an interface between the MAC and the PHY protocol layers and are defined by how the data on the channel is to be transmitted. A task of the MAC is to multiplex the channels so that they are correctly connected. Additionally the MAC takes care of all the scheduling. The scheduling determines what is to be transmitted in what re-source block, where a rere-source block is a certain frequency during a certain time period. All scheduling in the cells belonging to an eNodeB, both for the downlink and for the uplink, is performed in this eNodeB, and the necessary information is then transmitted in advance to the affected users. Similarly to the RLC layer, the MAC also has the ability to request retransmissions in the cases of erroneous detections of signals sent through the air.

The physical layer, as the lowest layer, is responsible for, among other things, cod-ing, modulation and mapping between transport channels and physical channels. LTE defines a scheduling entity called a frame, and a frame in turn consists of ten subframes. A subframe divides the available frequencies into resource blocks of 180kHz that each spans 1ms. The whole frame thus covers a time period of 10ms.

(36)

20 3 LTE Overview

Each physical channel corresponds to a set of these blocks, all data belonging to a certain physical channel may only be transmitted during these times and at the corresponding frequencies. Following the end of each 10ms frame there is another frame, repeating into infinity.

Some of the available physical channels are the Physical Downlink Shared Chan-nel (PDSCH), the Physical Downlink Control ChanChan-nel (PDCCH), the Physical Up-link Shared Channel (PUSCH), the Physical UpUp-link Control Channel (PUCCH) and the Physical Random-Access Channel (PRACH). The PRACH, PDSCH and PUSCH in particular are used in the random access procedure.

3.3 Random Access in LTE

The random access functionality in LTE has the purpose of establishing a connec-tion between a user and a cell. This is needed in the cases when there has been no prior communication between the cell and the user, since in that case the user and cell have no way of deciding a common schedule through which they can communicate. There are several situations where this is relevant, some examples are when a user’s equipment has just been turned on and needs to connect to the system, when a user needs to perform a handover operation as it is about to leave a cell and enter another and thus needs to connect to the new cell, and when an already existing connection is lost due to a previous failure in communication. An inherent issue with the random access procedure is that the user needs to acquire necessary information such as what frequencies to use and when to trans-mit its random access messages. In order to solve this, the cell broadcasts the required information which allows the user start the procedure.

The first part of the required information is received from a system informa-tion block (SIB). Several SIBs exist and they are transmitted on predetermined frequencies with certain time periods. The user simply has to repeatedly try to receive this message and it will eventually succeed in doing so. The second SIB in the set of SIBs provides information to the user of at which frequency and during what times, relative to the reception of the received downlink synch signal, the cell can receive the initial random access message.

A known and predetermined reference signal, often called a pilot, is also required and is therefore transmitted regularly by the cells. Since the user already knows the form and shape of the pilot the only knowledge gained by receiving this sig-nal is how the channel has affected the sigsig-nal. The channel refers to the radio link between the cell and the user. The user compares the received signal power with the power that it knows was used when transmitting the signal to attain an estimate of the channels path loss. This information is later used when the user transmits its own initial random access message to calculate the needed transmis-sion power.

(37)

3.3 Random Access in LTE 21

Two different kinds of random access procedures exist: contention-based and contention-free. As the name implies the contention-based scheme is what is used when several users simultaneously try to connect to the same cell using the same resources, this is the standard approach that is considered in this thesis. The contention-free scheme avoids collision between users by assigning a unique preamble, which is a kind of message used during the first step of the random access, to the user prior to the otherwise first step of the random access proce-dure. This is only possible when the cell somehow has information about the user beforehand, meaning that it is only possible in some use cases such as han-dover, and not in the initial connection case where the system has no information about the user whatsoever. The cost of doing a contention-free operation is the need for additional preambles; the system only has a certain amount of pream-bles, of which a majority is used for the contention-based access, so the use of the contention-free access is limited. The gain, on the other hand, is that it is possible to avoid the random delay caused by the risk of collision, when several users in-advertently use the same preamble at the same time causing the signals to collide and be indistinguishable at the receiver cell, in the contention-based scheme [11]. Figure 3.3 shows the set of messages sent in the contention-based procedure. The UE and eNodeB send a total of four messages, assuming that no step fails in which case the UE will have to start over. It is the UE that initiates the procedure.

Message 1 Message 3 Message 2 eNodeB User Message 4

Figure 3.3:The different messages sent during the contention-based random access procedure. The UE initiates the process by transmitting a preamble, after which the eNodeB and UE keep responding with consecutive messages until a total of four messages have been sent.

(38)

22 3 LTE Overview

3.3.1 Message 1

The message that initiates the random access process, assuming a contention-based scheme, is sent from the user and is called the random access preamble. The PRACH is used for this transmission. The purpose of the message is to in-form the cell that this user is trying to access the network and to enable the cell to prepare the transmission of the second message.

The user chooses the preamble from a set of at most 64 preambles, possibly a bit less as the number of preambles used for contention-free access is subtracted from this number. While the function that determines what preamble to use is de-terministic the input parameters to this function, such as the amount of data the user intends to transmit, can be considered to be randomly distributed [9]. This is equivalent to randomly choosing a preamble. Having decided upon a certain preamble the user then chooses the transmission power according to the equation

Pt= min{Pmax, PinitialTarget×P × Pramp× ∆}, (3.1)

where Ptis the transmission power, Pmaxis the maximum allowed power, PinitialTarget

is the baseline power value, P is the estimated path loss, Prampis a ramping factor

and ∆ is a constant value set by the system. If the user does not receive the ex-pected answer it then keeps retransmitting a preamble but with a higher power,

Prampcorresponds to this gradual increase in power and is increased according to

a set protocol with each failed transmission.

As the preamble that the user chooses is random there is a possibility that other users choose the same preamble. Since the preambles do not contain any user information a certain preamble will look the same no matter who sends it. If two or more users send the same preamble the result will therefore be that the cell receives the approximate sum of the two signals but perceives it as a single signal. The cell will then proceed as usual with message two which will likely lead to a conflict that needs to be resolved. Therefore it is not desirable that different users pick the same preamble at the same time.

3.3.2 Message 2

The second message in the random access procedure is the Random Access Re-sponse (RAR). The second message is used to inform the user of its relative timing error so that it will be able to synchronize with the cell, scheduling information on when and on what frequency the user may transmit the third message of the process, and a temporary identification for the user. The preamble signature is also included in the RAR so that the user can know whether this specific RAR is directed towards itself or another user. The transmission is done on the PDSCH [9].

In the event that there was a collision in the first step, that is if two users sent the same preamble at the same time, the cell will still only transmit a single RAR

(39)

3.3 Random Access in LTE 23

as it cannot distinguish between the two signals. Both users are then likely to suc-cessfully decode the incoming RAR and proceed to the third step with the same scheduling grant and identification.

The contention-free variant will stop at this step as the remaining steps are su-perfluous when it is already known that there was no collision.

3.3.3 Message 3

This message which is sent from the user through the PUSCH is the first of two steps in ensuring that there is no contention. The cell identifies the transmitter by the temporary identification assigned to the user during the second message. Additionally a kind of identification that is unique to the user is also part of this message as it is needed in order for the cell to be able to actually tell any contending users apart.

3.3.4 Message 4

The fourth and final message is sent on the PDSCH and contains information about which one user was the intended receiver of the grant given in the second message. It does this by choosing one out of possibly several unique identities received during step 3. If there was no contention then the single user will be identified as the chosen one, and if there was contention one user will be chosen. In both cases the chosen user will receive a message containing their unique iden-tification. Any contending users will upon receiving the same message realize that they were not the intended target of the message and will start over from the beginning of the procedure.

(40)

(41)

4

System Control Plane

The concept of the system control plane (SCP) was developed in order to solve some problems that have been identified in LTE and to improve the logical struc-ture of the upcoming 5G system. The lack of logical separation between the differ-ent kinds of signals in LTE causes a decreased ability to make use of the multiple input multiple output (MIMO) capabilities existing in the system. This is because some broadcasted system information has certain constraints in how the antenna resources are to be used, and these constraints are then also unnecessarily im-posed on all other kinds of signals as well. Another issue with LTE is that with time the amount of undefined resource elements has decreased as more and more of the resources are used for new features. Furthermore, the transmission of all mandatory reference signals prevent the power consuming components in the base station from entering sleep mode. The lack of undefined resource elements also makes it overall more difficult to extend the system with future changes. This chapter gives some background on the SCP and describes the problems, so-lution, motivations and the expected effects in more detail.

4.1 Definition and Proposed Implementation

The system control plane (SCP) is defined by the set of signals and procedures that are necessary for basic system functionality. This includes, but is not limited to, the the random access and paging procedures, as well as the signals necessary to execute these procedures. Random access is the procedure of a user making initial contact with a cell, and paging is the procedure of a cell locating and send-ing a message to a user.

Given this definition a SCP could be said to exist in LTE as well. Such a SCP

(42)

26 4 System Control Plane

would not be as useful, however, due to constraints specific to LTE. LTE for in-stance specifies that system information and data should be transmitted using the same antennas, meaning that you would not be able to logically separate the signals of the SCP from the other signals anyway. This prevents the realization of a logical separation of different kinds of transmissions, which is a desired prop-erty of the SCP.

Primarily we want the SCP to have two properties, the signals and procedures belonging to it should have an ultra-lean design, and as mentioned previously there should be a logical separation of transmissions belonging to the SCP from all other transmissions. As long as these properties exist, and the definition is adhered to, any implementation would be acceptable.

The implementation that we consider in this thesis contains two different signals, the access information table (AIT) and the system signature index (SSI). The AIT is a table that provides necessary system information to the user, including the information that is necessary for executing the random access procedure, and the SSI is an index to this table which indicates which part of the table is relevant. The SSI also fills the function of giving the users a signal that can be measured in order to calculate the power needed for transmitting the preamble in the ran-dom access procedure, something that in LTE was done by cell specific reference signals.

As for the procedures, we are only interested in the implementation of the ran-dom access procedure in this thesis. As mentioned earlier, the procedure here uses the SSI instead of cell specific reference signals to measure the power to use when transmitting the random access preamble. All other changes to the proce-dure, when compared to LTE, varies between simulations and are explained in Chapter 5. These changes consist of different ways of making use of several cells, both when receiving the first message of the procedure and when transmitting the second message of the procedure.

4.2 Static Signals

In LTE there are many different signals that are sent. The amount of different defined data channels described in Section 3.2 alone indicates that there are many types of information that passes between the mobile users and the eNodeBs. When analysing the potential areas of improvement for 5G, researchers have ob-served that a particular subset of the transmitted signals are responsible for a very large part of the total number of transmissions [5]. This subset consists of all the signals that must be sent regardless of user activity and system state. These signals are called static.

The static subset of the signals are used for the most basic system functional-ity and include broadcasted system information such as the AIT and the SSI. The

(43)

4.3 Dynamic Signals 27

AIT and SSI are always being broadcasted with a set periodicity. The system can not know when new users will appear so it always has to provide the information necessary for the users to be able to initiate random access.

4.3 Dynamic Signals

The set of dynamic signals is the complement of all static signals. The dynamic signals are only sent under certain conditions, when there is an explicit need for it. For example signals that contain data that a user has requested are dynamic. The SCP also includes dynamic signals. Both the random access procedure and the paging procedure are only executed under certain conditions, when a user needs to gain access to the network and when the network needs to inform a user of something. For this reason both of these procedures consist of dynamic signals.

4.4 Problems in LTE

In LTE there is a scheduler that allocates signals to the time-frequency resources according to the needs of the system. The scheduler assigns resources to the dynamic signals as they appear. The resulting schedule can be seen as a grid of squares, where each square signifies a specific transmission at a certain frequency and time.

The way LTE is designed ensures that the reference signals, described in Section 3.3, are sent as if the system is always operating at maximum capacity, mean-ing that signals are transmitted all the time and essentially all time-frequency resources are in use. This does of course guarantee a minimum of throughput at all times and simplifies the design. But considering the scarcity of the situations where the system is under heavy load and suffers from lack of time-frequency resources, it can also be viewed as a waste of resources. A consequence of this frequent use of reference signals is the inability for the cell to turn off and initi-ate sleeping mode. Since the time it takes to turn on the cell again when sleeping already takes up a large part of the time between reference signal transmissions there is not much left for actual sleep, which results in potentially more energy than necessary being used.

Figure 4.1 [10] shows how many empty subframes are used for increasingly high traffic scenarios. The figure represents scenarios of a country wide network, in-cluding both rural and urban areas, that are subject to different loads. Going from a scenario where there are no users present at all, the no traffic-scenario, to a scenario with a large amount of traffic, the high traffic-scenario, the number of empty subframes has only decreased by 19%. At first this appears to be a good result, considering that the more free resources we have the better. What causes

(44)

28 4 System Control Plane

these numbers, however, is that even when there are no users present in the sys-tem there is already a significant amount of subframes that are not empty. So many of the subframes are already used that adding large amounts of traffic only makes a marginal difference in the number of empty subframes. This indicates that much energy is spent in LTE even when there are no users present in the system.

No traffic Low traffic Medium traffic High traffic 0 10 20 30 40 50 60 70 80 90 100

Empty Subframe Ratio [%]

81% 90%

95%

Figure 4.1: The ratio of empty subframes relative to the total number of subframes for increasingly high traffic scenarios [10].

Figure 4.2 [10] shows how much energy is spent on static and dynamic trans-missions respectively. The scenario is the same as in Figure 4.1, a country wide network with varying amounts of traffic. When there is no traffic 100% of the energy is spent on static transmissions, whereas the high traffic scenario still has more than 90% of the energy spent on the static transmissions. This further sup-ports what Figure 4.1 indicated; even in high traffic scenarios a very large amount of the total energy is spent on static transmissions, of which reference signals are a very large part. The surprising distribution of energy suggests that it would be worthwhile to improve upon how static signals are handled in LTE.

In LTE it is also hard to truly utilize massive MIMO due to the very frequent ref-erence signals. To start with, it is hard to use MIMO for different frequencies at the same time, meaning that having fewer reference signals to transmit is bene-ficial. It is also troublesome to be able to handle both signals that are received weakly, such as broadcasted signals, and signals that are received strongly, such as those that are sent using for MIMO. This is so because it puts larger demands on the hardware used in the user equipment. While possible it can also be diffi-cult to have all antennas coordinate when you take into account the time it takes

(45)

4.4 Problems in LTE 29

No traffic Low traffic Medium traffic High traffic 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Area Power [kW/km 2] Static Energy Dynamic Energy 2.9% 7.4% 1.4%

Figure 4.2:The amount of energy used by the static (blue) and dynamic (red) for increasingly high traffic scenarios [10].

to change modes and cooperate. In the case of analogue beamforming it is easier to adjust the antennas properly if there is more uninterrupted time than what is currently available due to the frequent reference signals.

Another issue that has become prevalent is the cost of the increased interference between cells that is caused by having more dense networks. As the frequency spectrum is used for many more things than in the past it is convenient to attempt to use higher frequencies due to the lack of free low frequency bands. As shown by (2.2) however, using higher frequencies cause a higher path loss. Therefore to achieve the same coverage we need to make the system denser in the sense that the eNodeBs should be placed less sparsely. This also increases the interference forcing more signals to be retransmitted, thus leading to increased overhead. An-other reason for wanting to make the system denser is to increase performance. With more cells there would of course be less users per cell, leading to a lower load on each cell. The increased cost could have been avoided if the overhead did not scale with the increased system eNodeB density.

Yet another issue is the the drawbacks that come as a result of using more ad-vanced transmission techniques. One example of an underutilized technique is antenna tilting. Antenna tilting can be used at an eNodeB to redirect the antennas towards temporary hotspots and increase performance at that location. However, a side effect of this is that the coverage changes. It is easy to imagine the user dissatisfaction if some users were to lose coverage at critical times despite being at a place where there previously was coverage, and by extension it is not hard to see why the operators are unwilling to make use of this technique. Relative to

A Study on Random Access Performance in Next Generation Mobile Network Systems

Institutionen för systemteknik

Department of Electrical Engineering

Examensarbete

A Study on Random Access Performance in Next

Generation Mobile Network Systems

A Study on Random Access Performance in Next

Generation Mobile Network Systems

Examensarbete utfört i Kommunikationssystem

vid Tekniska högskolan vid Linköpings universitet

av

Sammanfattning

Abstract

Acknowledgments

Contents

Notation

1

Introduction

1.1

Motivation

1.2

Purpose

1.3

Problem Formulation

1.4

Limitations

1.5

Thesis Outline

2

Wireless Transmission Overview

2.1

Basics

2.2

Pre- and post-processing

2.2.1

Source coding

2.2.2

Channel coding

2.2.3

Modulation

2.3

Fading

2.4

Diversity

2.4.1

Spatial Diversity

2.4.2

Delay Diversity

3

LTE Overview

3.1

Cellular Networks Infrastructure

3.2

LTE Structure

3.3

Random Access in LTE

3.3.1

Message 1

3.3.2

Message 2

3.3.3

Message 3

3.3.4

Message 4

4

System Control Plane

4.1

Definition and Proposed Implementation

4.2

Static Signals

4.3

Dynamic Signals

4.4

Problems in LTE